International Journal of Systematic and Evolutionary Microbiology (2000), 50, 251–258 Printed in Great Britain

Lactobacillus mucosae sp. nov., a new species with in vitro -binding activity isolated from pig intestine

Stefan Roos,1 Fredrik Karner,1 Lars Axelsson2 and Hans Jonsson1

Author for correspondence: Hans Jonsson. Tel: j46 18 673382. Fax: j46 18 673392. e-mail: hans.jonsson!mikrob.slu.se

1 Department of A new species from pig small intestine has been identified. In an Microbiology, Swedish attempt to isolate Lactobacillus reuteri strains carrying the putative University of Agricultural Sciences, Box 7025, colonization-factor gene (mub, for mucus binding) a mub-derived gene probe SE-75007 Uppsala, was used to screen pig intestinal material. A number of isolates were obtained Sweden and primary characterization showed that they were Gram-positive, - 2 MATFORSK, Norwegian negative, non--forming, non-motile rods. Growth occurred at 45 SC but Food Research Institute, not at 15 SC and the DNA GMC content was 46 mol%. analysis Osloveien 1, N-1430 A/ s, Norway together with DNA–DNA hybridization and analysis of the 16S rRNA sequence revealed that the new isolates represent a previously undescribed Lactobacillus species closely related to L. reuteri, and Lactobacillus pontis. The name Lactobacillus mucosae is proposed for this species and the type strain is S32T.

Keywords: Lactobacillus mucosae, new species, mucus binding

INTRODUCTION immune stimulation and protection against certain forms of cancer (Elmer et al., 1996; Isolauri et al., The use of (LAB) in the production 1998; Lichtenstein & Goldin, 1998). One of the basic and preservation of food and animal feed dates back assumptions regarding the important features of pro- several thousand years. Today this use is manifested by biotic micro-organisms is the need for colonization various dairy products such as cheese and yoghurt, ability (Huis in’t Veld et al., 1994; Brassart & Schiffrin, fermented sausages, vegetables and silage for animal 1997). In contrast to the case of many pathogenic feed. Recently, a great deal of interest has been focused bacteria, little is known about the mechanisms by on some members of the LAB with regard to their use which LAB interact with host components in the as (Fuller, 1989; Marteau & Rambaud, intestinal tract. However, in recent years several 1993; Salminen et al., 1996). The term ‘’ reports have begun to establish a knowledge base on refers to live organisms that are administered to how lactobacilli adhere to the intestinal mucosa animals or humans via feed or food products and are (Adlerberth et al., 1996; Roos et al., 1996; Yamamoto in some way beneficial to health (Fuller, 1989). The et al., 1996; Granato et al., 1999). LAB that are currently used as probiotics are primarily species of Lactobacillus and Bifidobacterium. In this We have recently cloned and sequenced an extremely context, bifidobacteria are often included in the LAB large gene, mub, from the pig intestinal isolate Lacto- group since they have many features in common with reuteri strain 1063, encoding a cell-surface this group. However, in contrast to the LAB group in with mucus-binding activity (S. Roos & H. general, bifidobacteria belong to the high GjC group Jonsson, unpublished results). This protein, termed of Gram-positive bacteria. The positive effects Mub, has a molecular mass of 358 kDa and contains attributed to these probiotic organisms include two types of large amino acid repeats. The parental stabilization of the normal microflora, protection strain, 1063, has very good binding activity against against pathogens, lowering of cholesterol levels, crude pig mucus in vitro. Furthermore, both the native ...... and the recombinant Mub protein were shown to Abbreviations: LAB, ; RDP, Ribosomal Database interact with immobilized mucus material. In order to Project. clarify the correlation between the presence of mub and The GenBank/EMBL/DDBJ accession number for the 16S rDNA gene the ability to adhere to mucus material, we screened sequence of strain S32T is AF126738. pig small-intestinal mucosa for bacteria harbouring

01181 # 2000 IUMS 251 S. Roos and others this gene. In this work we describe the isolation of a determined using HPLC, as described by Mesbah et al. number of strains that are reactive with a gene probe (1989). derived from mub and that exhibit binding to mucus DNA–DNA hybridization. DNA–DNA hybridization was material in vitro. Further characterization of these performed at DSMZ, Braunschweig, Germany. DNA isolates showed that they represent a new Lactobacillus hybridization was carried out according to De Ley et al. species that is closely related to L. reuteri, Lactobacillus (1970) with modifications as described by Escara & Hutton fermentum and Lactobacillus pontis. Interestingly, (1980) and Huß et al. (1983), using a Gilford System 2600 three strains of lactobacilli previously isolated from spectrophotometer equipped with a Gilford 2527-R thermo- programmer and plotter. Renaturation rates were calculated pig small intestine (Axelsson & Lindgren, 1987; by the program : (Jahnke, 1992). Wadstro$ m et al., 1987) were shown to belong to this mub Colony hybridization. The original isolates were recultivated new species. These isolates also harbour and and single colonies were inoculated on Protran BA85 possess the ability to adhere to pig mucus in vitro.An nitrocellulose filters (Schleicher & Schuell) placed on top of oligonucleotide probe that can be used for rapid MRS-agar plates and grown overnight in anaerobic jars at identification of Lactobacillus mucosae sp. nov. is also 37 mC. Colony hybridization was performed according to described. Sambrook et al. (1989) with a modified protocol for lysis of the cells which includes a prolonged incubation with SDS METHODS (30 min). An internal 1100 bp restriction fragment from the mub gene of L. reuteri strain 1063 (S. Roos & H. Jonsson, $# Origin of the strains. Strains 1028, 1031 and 1035 were unpublished results) was labelled with P, with the Multi- isolated from pig small intestine and described earlier prime DNA labelling system (Amersham), and used as a (Axelsson & Lindgren, 1987; Wadstro$ m et al., 1987). Strains probe in the hybridizations. S5, S14, S15, S17 and S32T were isolated in this work. Small intestine from a newly slaughtered pig was collected from Western blot analysis. Bacteria were grown for 48 h at 37 mC the slaughterhouse. In the laboratory, the intestine was in 100 ml Lactobacillus Defined Medium II (Kotarski & sectioned and 15-cm-long sections were cut open and rinsed Savage, 1979) supplemented with 1% instead of . The cells were sedimented by centrifugation and the with ice-cold PBS (8n0 g NaCl, 0n2 g KCl, 1n44 g Na HPO ;2 # % growth medium was sterile-filtered through a 45 µm filter H#O and 0n2g KH#PO% per 1000 ml dH#O) in order to remove loosely associated intestinal material. Mucosal and dialysed against 1 mM EDTA overnight with several material was then released by gently scraping the intestine changes of buffer. The material was then lyophilized and with a spatula. The released material was collected in a tube dissolved in 1n5mldH#O. The protein preparations were with ice-cold PBS and subsequently diluted and used to mixed 1:1 with sample buffer containing SDS and mercapto- inoculate agar plates. L. reuteri DSM 20016T, L. fermentum ethanol and separated by SDS-PAGE. The electrophoresis DSM 20052T and L. pontis DSM 8475T were obtained and staining of the gel with Coomassie blue was performed from the Deutsche Sammlung von Mikroorganismen und with the PhastSystem (Pharmacia Biotech) according to the Zellkulturen (DSMZ), Braunschweig, Germany, and Lacto- manufacturer’s instructions. The were blotted to a bacillus acidophilus ATCC 4356T was obtained from the Hybond-C nitrocellulose membrane (Amersham Life Sci- American Type Culture Collection (ATCC), Manassas, VA, ence) by diffusion at 65 mC for 45 min. The membrane was USA. blocked in PBST (PBS supplemented with 0n05% Tween 20, pH 7 3) for 1 h at 37 C and then incubated overnight at 4 C n " m m Culture conditions. Primary isolation was done on Lacto- with 10 µgml− antibodies against extracellular proteins bacillus Selective Agar (BBL) in anaerobic jars under a from L. reuteri strain 1063 (S. Roos & H. Jonsson, un- CO#jN# atmosphere (GasPak System, BBL) at 37 mC. All published results) which had been preadsorbed with an further cultivation was done at 37 mC on De Man– Escherichia coli lysate according to Sambrook et al. (1989). Rogosa–Sharpe (MRS) agar (Oxoid), in anaerobic jars or in After being washed with PBST, the membrane was incubated MRS broth (Oxoid) unless stated otherwise. with HRP-conjugated goat anti-rabbit IgG (Bio-Rad), Morphological characteristics. Cell morphology was diluted 1:1000, at 37 mC for 1 h. After being washed, the observed using phase-contrast microscopy. Gram deter- membranes were finally developed with 4-chloro-1-naphthol mination was performed using both Gram-staining and the as substrate. KOH method of Gregersen (1978). Assay of mucus binding. Mucus was prepared from the same Physiological and chemical characterization. Sugar-fermen- intestine that was used for isolation of bacteria. By scraping tation patterns were determined using the API 50 CH system the inside of the intestine with a spatula, material was (BioMe! rieux). Cell wall analysis was performed at DSMZ. removed and collected in 200 ml ice-cold PBS. The resulting The preparation of cell walls and the determination of suspension was centrifuged first at 11000 g for 10 min and structure were carried out using the methods then at 26000 g for 15 min in order to remove cells and described by Schleifer & Kandler (1972), with the modi- particulate matter. The crude mucus preparation was stored fication that TLC was performed on cellulose sheets instead at 20 mC. The mucus material was diluted to approximately −" of paper. The lactic acid configuration was determined using 100 µgml in 50 mM Na#CO$ buffer, pH 9n7, and incu- a commercial test kit (Boehringer Mannheim). Arginine bated overnight in microtiter wells (Greiner) (150 µl per cleavage was determined using the method of Collins & Lyne well) at 4 mC with slow rotation. The wells were blocked with (1970). Catalase activity was determined by transferring PBS with 1% Tween 20 for 1 h and thereafter washed with fresh colonies from MRS agar to a glass slide and adding PBST. The bacteria were grown in MRS broth for 16 h at 5% H#O#. 37 mC, washed once in PBST and diluted to OD'!! l 0n5 DNA base composition. The GjC content of the DNA (1 cm cuvette; Beckman DU650) in the same buffer. Bac- (mol%) was determined at DSMZ. The DNA was isolated terial suspension (150 µl) was added to each well and by chromatography on hydroxyapatite using the procedure incubated for 1 h at room temperature. The wells were of Cashion et al. (1977). The GjC content of the DNA was washed with PBST and binding was examined with an

252 International Journal of Systematic and Evolutionary Microbiology 50 Lactobacillus mucosae sp. nov. inverted microscope. The buffer was poured off and, after derived from this gene from L. reuteri strain 1063 for the wells had dried, the OD%&! was measured in an ELISA screening lactobacilli isolated from pig small-intestine plate reader (EL309 Autoreader; BIO-TEK Instruments). mucosa. Approximately 100 colonies were screened 16S rRNA gene sequencing. The almost complete 16S rRNA with colony hybridization, of which five (S5, S14, S15, gene was amplified by PCR by using slightly modified S17 and S32T) reacted with the probe. Fifty additional domain Bacteria-specific primers according to Weizenegger Lactobacillus strains previously isolated from pig small et al. (1992). The primer sequences were 5h-AGAGTTTG- intestine [and characterized by Wadstro$ m et al. (1987) ATYMTGGC-3h and 5h-AGAAAGGAGGTGATCC-3h. and Axelsson and Lindgren (1987)] were also screened PCR reactions involved 35 cycles under the following with the probe: three of them (1028, 1031 and 1035) conditions: 94 mC for 30 s, 54 mC for 30 s and 72 mC for 80 s; the resulting PCR products were purified from agarose gels. gave positive signals. Strains reacting with the probe Both strands of the purified fragments were sequenced using were recultivated from the master plate and checked the Thermo Sequenase dye terminator cycle sequencing pre- by SDS-PAGE and Western blotting for production of mix (Amersham) and the automated sequence analyser ABI the mucus-binding protein, Mub. Since this protein is PRISM 377XL (Perkin Elmer). The same primers that were very large (358 kDa), bands located at the top of the used for the amplification were used (together with ad- gel and reactive with antibodies against extracellular ditional customized internal primers) to sequence the PCR proteins from L. reuteri strain 1063 were considered to products. be Mub or a related protein. All strains that were Phylogenetic analysis of the sequence data. The sequences reactive with the mub probe also produced a large achieved from the new isolates were used for searching in the protein reacting with the antibodies. The same set of public databases (GenBank and the Ribosomal Database strains was also positive in the assay of mucus binding Project, RDP). The sequences representing the best matches (data not shown). were retrieved and aligned using the   program (Thompson et al., 1994). The sequences were manually modified before the alignment and approximately 1450 Colony and cell morphology nucleotides covered by all sequences were used. The fol- T lowing sequences from the type strains of the respective Colonies of strains S5, S14, S15, S17, S32 , 1028, 1031 species were used: L. fermentum, M58819; L. reuteri, and 1035 were white, smooth and convex. After X76328; L. pontis, X76329; Lactobacillus vaginalis, X61136; anaerobic growth for 2 d on MRS agar at 37 mC, the Lactobacillus oris, X94229; Lactobacillus panis, X94230; colonies were 1–2 mm in diameter. Cells of all strains Lactobacillus sakei, M58829; Lactobacillus casei, X61135; were non-spore-forming, non-motile rods 1i2–4 µm Lactobacillus brevis, X61134; Lactobacillus salivarius subsp. in size. The Gram-reaction was positive. salivarius, AF089108; Lactobacillus delbrueckii subsp. delbrueckii, M58814; Lactobacillus fructivorans, X76330; DNA base composition Pediococcus acidilactici, M58833. A distance matrix was T calculated with the  program of the  package The mean GjC content of strain S32 DNA, based (Felsenstein, 1993), using the Kimura 2-parameter model, on three determinations, was 46n5p0n2 mol%. The and a phylogenetic tree was constructed with the  GjC content of strains 1028, 1031 and 1035 was program ( package) using the neighbour-joining determined previously (Axelsson & Lindgren (1987): method. The statistical significance of the grouping was estimated by bootstrapping (100 replicates) using the the values for these three strains were 49, 47 and programs , ,  and  (all 49 mol%, respectively. from the  package). Hybridization with a specific rDNA-targeted probe. Bacteria DNA–DNA hybridization were grown and lysed on filters as described for the colony On the basis of the 16S rRNA sequence analysis, three hybridization. The oligonucleotide LM16, with the sequence reference strains were chosen for DNA–DNA hy- 5h-GTAAACCAACGTCAAGTCC-3h, was designed to be T specific for 16S rRNA and 16S rDNA from the new isolates bridization experiments with strain S32 . These were L. reuteri DSM 20016T, L. fermentum DSM 20052T by comparing sequences from different species. The - T  program at RDP was used to confirm the specificity and L. pontis DSM 8475 . The results showed hom- of the probe. The customized oligonucleotide was labelled ology values of 40 5, 41 5 and 52 2%, respectively, $# n n n with P using T4 polynucleotide kinase, according to between these strains and L. mucosae S32T. Sambrook et al. (1989). The prehybridization, hybridization et al and washing were performed according to Sambrook . Physiological and chemical characterization (1989). The temperature during hybridization and washing was 42 mC. Analysis of the cell wall of L. mucosae strain S32T revealed the presence of ornithine and aspartic acid, RESULTS which is consistent with an Orn--Asp peptidoglycan Identification of colonies reactive with the mub type. All strains grew well at 45 mC but not at 15 mC. probe, producing Mub protein and possessing Gas was produced from glucose. All strains grew well mucus-binding activity in liquid and on solid MRS media in anaerobic jars. Weak growth also occurred on solid MRS media in the In an attempt to isolate L. reuteri strains carrying the presence of air. Arginine was cleaved by all strains. gene for the mucus-binding protein (Mub) (S. Roos & Catalase activity was negative in all strains. -Lactate H. Jonsson, unpublished results), we used a probe and -lactate were produced by all strains. The

International Journal of Systematic and Evolutionary Microbiology 50 253 S. Roos and others

Table 1. Differential characteristics of Lactobacillus mucosae sp. nov. and closely related lactobacilli ...... , Not determined; v, variable reaction.

Strain Utilization of:

NH3 from Growth at Gas from Glucose L-Arabinose D-Xylose Galactose arginine 15/45 mC glucose

L. mucosae S5 jk\jjjjjjj L. mucosae S14 jk\jjjjkkk L. mucosae S15 jk\jj j v kjk L. mucosae S17 jk\jjjjkjj L. mucosae S32T jk\jjjjkjj L. mucosae 1028 jk\jjjjjjj L. mucosae 1031 jk\jjjjjjk L. mucosae 1035 jk\jjjjjjk L. reuteri DSM jk\jjjjjkj 20016T L. fermentum DSM jk\jjjjkkj 20052T L. pontis DSM 8475T† jj\jkkjkkk L. oris DSM 4864TR kk\jj j¶ jj j j¶ L. panis DSM 6035TR kk\jj j j j j  L. vaginalis DSM jk\jj j¶ kk k j¶ 583TR

Strain Utilization of: Peptidoglycan GjC type content D-Fructose Aesculin Lactose Melibiose D-Raffinose Gluconate (mol%)

L. mucosae S5 kjkj j j   L. mucosae S14 kjkk k k   L. mucosae S15 kjkj j j   L. mucosae S17 kjkj j j   L. mucosae S32T kjkj j jOrn--Asp 46n5p0n2 L. mucosae 1028 v jjj j j  49* L. mucosae 1031 kjkj j j  47* L. mucosae 1035 kjkj j j  49* L. reuteri DSM jkjj j j‡ Lys--Asp 40–42‡ 20016T L. fermentum DSM jjjj j jOrn--Asp 52–54‡ 20052T L. pontis DSM 8475T† jkkk j jOrn--Asp 53§ L. oris DSM 4864TR jjjj j jLys--Asp 49–51F L. panis DSM 6035TR jjjj j jLys--Asp 48 L. vaginalis DSM jjjj j kOrn--Asp 38–41** 583TR

* Data from Axelsson & Lindgren (1987). † Data from Vogel et al. (1994). ‡ Data from Kandler & Weiss (1986). Data are typical for the species, not specifically the type strain. § The values for L. pontis LTH 2585 and 2586 (Vogel et al., 1994). R Data from Wiese et al. (1996). ¶ Data from Hammes et al. (1992). F Data from Farrow & Collins (1988). ** Data from Embley et al. (1989).

254 International Journal of Systematic and Evolutionary Microbiology 50 Lactobacillus mucosae sp. nov.

Relevant characteristics of the L. mucosae strains are summarized in Table 1.

16S rRNA sequence and phylogenetic analysis Almost the complete 16S rDNA sequence was de- termined for strains S14 and S32T, while partial sequences from each end of the gene were determined for strains 1028, 1031, 1035, S5, S15 and S17. Analysis of all sequences from the 5h- and 3h ends of the genes showed no differences except for a small number of ambiguities, so we concluded from these data that all isolates belong to the same species. The complete sequences from S14 and S32T were aligned and found to be identical. The complete S32T sequence was used to search the RDP and the highest similarity rank was found with L. reuteri, L. pontis and L. fermentum, the values being 95n1, 94n6 and 94n4%, respectively. The new sequence was then aligned with the sequences from members of the L. reuteri–L. fermentum branch of the L. casei–Pediococcus group and other repre- sentatives of this group and the alignment was used for a tree analysis. This analysis confirmed the close relationship of L. mucosae to L. reuteri and L. fermentum (Fig. 1).

Specific detection of L. mucosae with the aid of a rDNA probe Colony hybridization of the oligonucleotide probe LM16 (5h-GTAAACCAACGTCAAGTCC-3h) with all eight L. mucosae strains, L. reuteri DSM 20016T, L...... fermentum DSM 20052T and L. acidophilus ATCC Fig. 1. Unrooted phylogenetic tree, derived from 16S rRNA 4356T was performed. The oligonucleotide was selec- sequence analysis, showing the relationship between Lacto- bacillus mucosae strain S32T and members of the Lactobacillus tive for the L. mucosae strains under the conditions casei–Pediococcus group of lactobacilli. The sequence from used. Lactobacillus delbrueckii was used as an outgroup represen- tative. All strains are type strains. Approximately 1450 nucleo- tides from each sequence were used for the alignment. Bar, 1% DISCUSSION estimated sequence divergence. The genus Lactobacillus currently includes more than 60 species; three of these were described during 1998 (Morlon-Guyot et al., 1998; Edwards et al., 1998; Bohak et al., 1998) and one was described at the beginning of 1999 (Falsen et al., 1999). Members of following sugars were fermented: glucose, ribose, this genus have been isolated from a large number of maltose, saccharose, -xylose (7 of 8 strains), melibiose habitats, including oral and genital sites and the (7 of 8), -raffinose (7 of 8), gluconate (7 of 8), - gastrointestinal tracts of animals and humans. In pigs, arabinose (4 of 8), galactose (4 of 8) and lactose (1 of a number of Lactobacillus species have been iden- 8). There was no of -fructose, glycerol, tified and L. reuteri, L. fermentum, L. acidophilus, L. erythritol, -xylose, adonitol, -mannose, -sorbose, delbrueckii and L. salivarius are commonly reported as rhamnose, dulcitol, inositol, mannitol, sorbitol, methyl being isolated from pig intestine (Stewart, 1997). On α--glucoside, N-acetylglucosamine, salicine, cello- two different occasions separated by a 12-year interval, biose, trehalose, inuline, melizitose, amidon, glycogen, we have isolated strains of lactobacilli from pig small xylitol, -turanose, -lyxose, -tagatose, -fucose, intestine. Many of these isolates are L. reuteri and one -fucose, -arabitol, 2-keto-gluconate or 5-keto- of them, strain 1063, has been characterized with gluconate. Aesculin was hydrolysed by all strains. The respect to autoaggregation and mucus-binding ability minor differences in the sugar-fermentation patterns of (Roos et al., 1999; S. Roos & H. Jonsson, unpublished strains 1028, 1031 and 1035 found when data from this results). Three isolates (1028, 1031 and 1035) were work were compared with data presented by Axelsson partially described by Axelsson & Lindgren (1987) and & Lindgren (1987) cannot be explained at this point. these three strains, together with five new isolates (S5,

International Journal of Systematic and Evolutionary Microbiology 50 255 S. Roos and others

S14, S15, S17 and S32T), have been further investigated than one Lactobacillus species and the correlation of in this work. The common characteristic of these the presence of this gene with mucus-binding ability isolates, which originally attracted our attention, was argue that it represents an important part of the the reactivity with a gene probe derived from the colonization mechanism for lactobacilli in the pig putative colonization-factor gene mub from L. reuteri intestine. strain 1063 (S. Roos & H. Jonsson, unpublished results). When we further examined these isolates we also found that they exhibited in vitro characteristics Description of Lactobacillus mucosae sp. nov. similar to those of strain 1063, with respect to mucus- Lactobacillus mucosae (muhco.sae. L. gen. n. mucosae binding activity. Primary characterization using sugar- of mucosa). fermentation analysis indicated that these isolates were not L. reuteri and we therefore used nucleic acid Cells are Gram-positive, non-motile, non-spore- methods to investigate their taxonomic position in forming, catalase-negative rods with dimensions of more detail. Thus, 16S rRNA sequence analysis, GjC 1i2–4 µm. The cells occur singly, in pairs or as short mol% determination and DNA–DNA hybridization, chains and have an Orn--Asp peptidoglycan type. together with cell wall analysis, showed that they After anaerobic growth at 37 mC for 2 d, colonies on represent a previously undescribed species. On the MRS agar are 1–2 mm in diameter. They are white, basis of phylogenetic analysis, the new species can be smooth and convex. The cells are obligately hetero- allocated to the L. casei–Pediococcus group described fermentative and produce - and -lactic acid. They by Collins et al. (1991). The closest relatives are L. grow at 45 mC but not at 15 mC. Acid is produced from reuteri, L. fermentum and L. pontis, which show 16S glucose, ribose, maltose and saccharose. The majority rRNA sequence similarity values of 95n1, 94n4 and of strains also ferment -xylose, melibiose, -raffinose 94n6%, respectively. The construction of a phylo- and gluconate. -Arabinose, galactose or lactose may genetic tree confirmed the close affiliation of the new be utilized by some strains. -Fructose, glycerol, isolates with the L. reuteri–L. fermentum group. erythritol, -xylose, adonitol, -mannose, -sorbose, DNA–DNA hybridization of total DNA from S32T rhamnose, dulcitol, inositol, mannitol, sorbitol, methyl with the type strains of the three closely related species α--glucoside, N-acetylglucosamine, salicine, cello- showed an overall sequence homology in the range biose, trehalose, inuline, melizitose, amidon, glycogen, 40n5–52n2%, clearly indicating that L. mucosae xylitol, -turanose, -lyxose, -taganose, -fucose, constitutes a new species separate from these others. -fucose, -arabitol, 2-keto-gluconate and 5-keto- Our own data and also work by others have shown gluconate are not utilized. Aesculin is hydrolysed. that L. reuteri and L. mucosae can be found in the same Arginine is cleaved. The DNA GjC content of strain T part of the pig intestine. L. fermentum, though fre- S32 is 46n5p0n2 mol%. All strains of this species so quently reported as being present in the pig intestinal far described have a homologue to the mucus-binding tract, was not found in this niche by Axelsson & protein, Mub, and exhibit mucus-binding ability in Lindgren (1987). The species L. pontis, L. oris, L. vitro. vaginalis and L. panis have never been reported as colonizing this environment. However, it must be mentioned that few, if any, ecological studies on the ACKNOWLEDGEMENTS gastrointestinal-tract microbiota have employed 16S The authors thank Dr Norbert Weiss, DSMZ (Deutsche rDNA sequence determination for identification of the Sammlung von Mikroorganismen und Zellkulturen, bacteria. The methods that have been used in most Braunschweig, Germany) for advice regarding phylogeny studies either fail to detect all types of lactobacilli or and nomenclature. This study was supported by grants from fail to distinguish between certain species. The Swedish National Board for Industrial and Technical Development (NUTEK; P5974) in a co-operative project The relationship between L. reuteri and L. mucosae is with BioGaia Biologics AB (Stockholm, Sweden). further accentuated by the presence of the mub gene. Interestingly, the GjC content (mol%) of this gene is close to the overall G C content of L. mucosae and REFERENCES j ! rather different from that of other genes from L. Adlerberth,/ I., Ahrne, S., Johansson, M.-L., Molin, G., Hanson, reuteri. Although five of the isolates (S5, S14, S15, S17 L. A. & Wold, A. E. (1996). A mannose-specific adherence mech- and S32T) were primarily selected for harbouring the anism in Lactobacillus plantarum conferring binding to the mub gene, this was not the case with the three human colonic cell line HT-29. Appl Environ Microbiol 62, previously isolated strains (1028, 1031 and 1035) yet 2244–2251. the gene was still present. Thus, it seems that strains of Axelsson, L. & Lindgren, S. (1987). Characterization and DNA L. mucosae commonly carry this gene. In contrast, the homology of Lactobacillus strains isolated from pig intestine. presence of mub is restricted to a limited number of J Appl Bacteriol 62, 433–440. strains of L. reuteri. Altogether, this could implicate L. Bohak, I., Back, W., Ehrmann, M., Ludwig, W. & Schleifer, K. H. mucosae as the source of mub and L. reuteri as a (1998). Lactobacillus amylolyticus sp. nov., isolated from beer recipient of the gene at some point during the course of malt and beer wort. Syst Appl Microbiol 21, 360–364. evolution. The presence of a mub homologue in more Brassart, D. & Schiffrin, E. J. (1997). The use of probiotics to

256 International Journal of Systematic and Evolutionary Microbiology 50 Lactobacillus mucosae sp. nov. reinforce mucosal defence mechanisms. Trends Food Sci Microbiology and Functional Aspects, 2nd edn, pp. 255–268. Technol 8, 321–326. Edited by S. Salminen & A. von Wright. New York: Marcel Cashion, P., Holder-Franklin, M. A., McCully, J. & Franklin, M. Dekker. (1977). A rapid method for the base ratio determination of Jahnke, K.-D. (1992). BASIC computer program for evaluation bacterial DNA. Anal Biochem 81, 461–466. of spectroscopic DNA renaturation data from GILFORD Collins, C. H. & Lyne, P. M. (1970). Microbiological Methods. SYSTEM 2600 spectrophotometer on a PC\XT\AT type London: Butterworth. personal computer. J Microbiol Methods 15, 61–73. Kandler, O. & Weiss, N. (1986). Genus Lactobacillus Beijerinck Collins, M. D., Rodrigues, U., Ash, C., Aguirre, M., Farrow, J. A. E., AL Martinez-Murcia, A., Phillips, B. A., Williams, A. M. & Wallbanks, 1901, 212 .InBergey’s Manual of Systematic Bacteriology, S. (1991). Phylogenetic analysis of the genus Lactobacillus and vol. 2, pp. 1209–1234. Edited by P. H. A. Sneath, N. S. Mair, related lactic acid bacteria as determined by reverse trans- M. E. Sharpe & J. G. Holt. Baltimore: Williams & Wilkins. criptase sequencing of 16S rRNA. FEMS Microbiol Lett 77, Kotarski, S. F. & Savage, D. C. (1979). Models for study of the 5–12. specificity by which indigenous lactobacilli adhere to murine De Ley, J., Cattoir, H. & Reynaerts, A. (1970). The quantitative gastric epithelia. Infect Immun 63, 1698–1702. measurement of DNA hybridization from renaturation rates. Lichtenstein, A. H. & Goldin, B. R. (1998). Lactic acid bacteria and Eur J Biochem 12, 133–142. intestinal drug and cholesterol metabolism. In Lactic Acid Edwards, C. G., Haag, K. M., Collins, M. D., Hutson, R. A. & Bacteria: Microbiology and Functional Aspects, 2nd edn, pp. Huang, Y. C. (1998). Lactobacillus kunkeei sp. nov.: a spoilage 269–277. Edited by S. Salminen & A. von Wright. New York: organism associated with grape juice . J Appl Marcel Dekker. Microbiol 84, 698–702. Marteau, P. & Rambaud, J.-C. (1993). Potential of using lactic acid bacteria for therapy and immunomodulation in man. Elmer, G. W., Surawicz, C. M. & McFarland, L. V. (1996). Bio- FEMS Microbiol Rev 12 therapeutic agents: a neglected modality for the treatment and , 207–220. prevention of selected intestinal and vaginal infections. JAMA Mesbah, M., Premachandran, U. & Whitman, W. (1989). Precise (J Am Med Assoc) 275, 870–876. measurement of the GjC content of deoxyribonucleic acid by high performance liquid chromatography. Int J Syst Bacteriol Embley, T. M., Faquir, N., Bossart, W. & Collins, M. D. (1989). 39, 159–167. Lactobacilus vaginalis sp. nov. from the human vagina. Int J Syst Bacteriol 39, 116–118. Morlon-Guyot, J., Guyot, J. P., Bot, B., de Haut, J. & Raimbault, M. (1998). Lactobacillus manihotivorans sp. nov., a new starch- Escara, J. F. & Hutton, J. R. (1980). Thermal stability and hydrolysing lactic acid bacterium isolated during cassava sour renaturation of DNA in dimethyl-sulphoxide solutions: ac- starch fermentation. Int J Syst Bacteriol 48, 1101–1109. celeration of renaturation rate. Biopolymers 19, 1315–1327. $ $ ! ! Roos, S., Aleljung, P., Robert, N., Lee, B., Wadstrom, T., Lindberg, Falsen, E., Pascual, C., Sjoden, B., Ohlen, M. & Collins, M. D. M. & Jonsson, H. (1996). A collagen binding protein from (1999). Phenotypic and phylogenetic characterization of a novel Lactobacillus reuteri is part of an ABC transporter system? Lactobacillus species from human sources: description of FEMS Microbiol Lett 144, 33–38. Lactobacillus iners sp. nov. Int J Syst Bacteriol 49, 217–221. Roos, S., Lindgren, S. & Jonsson, H. (1999). Autoaggregation of Farrow, J. A. E. & Collins, M. D. (1988). Lactobacillus oris sp. nov. Lactobacillus reuteri is mediated by a putative DEAD-box from the human oral cavity. Int J Syst Bacteriol 38, 116–118. helicase. Mol Microbiol 32, 427–436. Felsenstein, J. (1993).  (Phylogeny Inference Package) Salminen, S., Isolauri, E. & Salminen, E. (1996). Clinical use for version 3.5c. Department of Genetics, University of probiotics for stabilising the gut mucosal barrier: successful Washington, Seattle, WA, USA. strains and future challenges. Antonie Leeuwenhoek 70, Fuller, R. (1989). Probiotics in man and animals. J Appl Bacteriol 347–358. 66, 365–378. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Granato, D., Perotti, F., Masserey, I., Rouvet, M., Golliard, M., Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, Servin, A. & Brassart, D. (1999). Cell surface-associated lipo- NY: Cold Spring Harbor Laboratory. teichoic acid acts as an adhesion factor for attachment of Schleifer, K. H. & Kandler, O. (1972). Peptidoglycan types of Lactobacillus johnsonii La1 to human enterocyte-like Caco-2 bacterial cell walls and their taxonomic implications. Bacteriol cells. Appl Environ Microbiol 65, 1071–1077. Rev 36, 407–477. Gregersen, T. (1978). A rapid method for distinction of Gram- Stewart, C. S. (1997). Microorganisms in hindgut fermentors. In negative from Gram-positive bacteria. Eur J Appl Microbiol Gastrointestinal Microbiology, vol. 2, pp. 142–186. Edited by Biotechnol 5, 123–127. R. I. Mackie, B. A. White & R. E. Isaacson. New York: Hammes, W. P., Weiss, N. & Holzapfel, W. (1992). The genera Chapman & Hall. Lactobacillus and Carnobacterium.InThe Prokaryotes, 2nd Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994).  edn, pp. 1535–1594. Edited by A. Balows, H. G. Tru$ per, M. : improving the sensitivity of progressive multiple sequence Dworkin, W. Harder & K.-H. Schleifer. New York: Springer. alignment through sequence weighting, position-specific gap Huis in’t Veld, J. H. J., Havenaar, R. & Marteau, P. (1994). penalties and weight matrix choice. Nucleic Acids Res 22, Establishing a scientific basis for probiotic R&D. Trends 4673–4680. $ Biotechnol 12, 6–8. Vogel, R. F., Bocker, G., Stolz, P. & 7 other authors (1994). Huß, V. A. R., Festl, H. & Schleifer, K. H. (1983). Studies on the Identification of lactobacilli from sourdough and description of spectrophotometric determination of DNA hybridization from Lactobacillus pontis sp. nov. Int J Syst Bacteriol 44, 223–229. $ renaturation rates. Syst Appl Microbiol 4, 184–192. Wadstrom, T., Anderson, K., Sydow, M., Axelsson, L., Lindgren, S. Isolauri, E., Salminen, E. & Salminen, S. (1998). Lactic acid & Gullmar, B. (1987). Surface properties of lactobacilli isolated bacteria and immune modulation. In Lactic Acid Bacteria: from the small intestine of pigs. J Appl Bacteriol 62, 513–520.

International Journal of Systematic and Evolutionary Microbiology 50 257 S. Roos and others

Weizenegger, M., Neumann, M., Stackebrandt, E., Weiss, N. & Lactobacillus panis sp. nov., from sourdough with a long Ludwig, W. (1992). Eubacterium alactolyticum phylogenetically fermentation period. Int J Syst Bacteriol 46, 449–453. groups with Eubacterium limosum, Acetobacterium woodii and Yamamoto, K., Miwa, T., Taniguchi, H., Nagano, T., Shimamura, Clostridium barkeri. Syst Appl Microbiol 15, 32–36. K., Tanaka, T. & Kumagai, H. (1996). Binding specificity of Wiese, B. G., Strohmar, W., Rainey, F. A. & Diekmann, H. (1996). Lactobacillus to glycolipids. Biochem Biophys Res 228, 148–152.

258 International Journal of Systematic and Evolutionary Microbiology 50